Sedative-analgesic treatment of patients on long-term artificial ventilation aims at protection from stress related to their disease or therapy. By stabilising both the patient's vital functions and psychological state this treatment may contribute to therapeutic success. The choice of drugs depends primarily on the nature and course of the underlying disease. Midazolam and propofol are available as hypnotics for short-term sedation during the post-operative period. The purpose of this study was to evaluate the effects of both agents on cardiovascular function, cortisol production, lipometabolism, and the recovery period following 24-h sedation. METHODS. Twenty female patients (mean body weight: 72 kg, mean age: 60 years) were randomly assigned to receive either midazolam or propofol over 24 h following major abdominal surgery. Balanced anaesthesia (halothane/O2/N2O/fentanyl) was administered for the surgical procedure. Assisted ventilation was used in all patients during the post-operative sedation period. Sedation depth was maintained at III-IV on the Ramsey scale. On arrival in the intensive care unit (ICU), an initial i.v. bolus of midazolam 0.1 mg/kg or propofol 1 mg/kg was followed by a continuous infusion (midazolam: 0.1 mg/kg.h; propofol: 2 mg/kg.h). Supplementary boluses of one-half the initial dose were given if required. Post-operative analgesia was achieved with 3 mg intravenous piritramide at 2-h intervals. A 7F Swan-Ganz catheter was inserted in the pulmonary artery and haemodynamic and biochemical parameters were monitored at 4-h intervals over 24 h starting 2 h after arrival in the ICU. Catecholamines were measured by high-pressure lipid chromatography (HPLC), cortisol by radioimmunoassay, midazolam by HPLC and ultraviolet detection, and propofol by HPLC and fluorescence detection. Data were calculated as means. The statistical analysis was performed according to the Mann-Whitney test, and significance was accepted for P less than 0.05. RESULTS. On administration of the propofol bolus at the onset of sedation, a decrease in blood pressure was particularly observed in patients with masked hypovolaemia, however, this decrease was easily controlled by volume administration. Independent of the type of sedation, the haemodynamic parameters remained unchanged throughout the observation period. At all times of measurement the mean heart rate was lower in the propofol group (90/min) when compared with the midazolam group (100/min), however, this difference did not reach significance. There were also no significant differences in cardiac index at all times of measurement, although it increased in both groups within the first 12 h by 0.6 l/min.min2. In both groups this increase was associated with a reduction in peripheral resistance and an increase in rectal temperature. To achieve the desired sedation depth, midazolam was administered at a mean dosage of 0.11 mg/kg.h and propofol at 1.9 mg/kg.h. Catecholamine levels decreased in both groups within the first 8 h: after 8 h of sedation the plasma levels of noradrenaline and adrenaline were 525 and 65 pg/ml, respectively, in the midazolam group and 327 and 51 pg/ml in the propofol group. (ABSTRACT TRUNCATED AT 400 WORDS)